Pavel R. Hrma

Chemically durable iron phosphate glass wasteforms

Up to 40 wt% of a simulated high level waste, whose major components were 54.6 wt% Na2O, 14.9 wt% P2O5 and 8.3 wt% Fe2O3, was successfully vitrified into iron phosphate wasteforms whose chemical durability was equivalent to that of borosilicate glass wasteforms. Because of their high fluidity, the iron phosphate wasteforms could be melted in as little as 30 min at temperatures between 1015°C and 1200°C. The addition of 3–7 wt% CaF2 to the batch decreased the melting time and temperature, by as much as 100°C, and improved the chemical durability, especially for crystallized iron phosphate wasteforms. Iron phosphate wasteforms are concluded to be a practical alternative for vitrifying those nuclear wastes not well suited for borosilicate glasses.

Nepheline Precipitation in High-Level Waste Glasses : Compositional Effects and Impact on the Waste Form Acceptability

The impact of crystalline phase precipitation in glass during canister cooling on chemical durability of the waste form limits waste loading in glass, especially for vitrification of certain high-level waste (HLW) streams rich in Na 2 O and Al 2 O 3 . This study investigates compositional effects on nepheline precipitation in simulated Hanford HLW glasses during canister centerline cooling (CCC) heat treatment. It has been demonstrated that the nepheline primary phase field defined by the Na 2 O-Al 2 O 3 -SiO 2 ternary system can be used as an indicator for screening HLW glass compositions that are prone to nepheline formation. Based on the CCC results, the component effects on increasing nepheline precipitation can be approximately ranked as Al 2 O 3 > Na 2 O > Li 2 O ≈ K 2 O ≈ Fe 2 O 3 > CaO > SiC 2 . The presence of nepheline in glass is usually detrimental to chemical durability. Using x-ray diffraction data in conjunction with a mass balance and a second-order mixture model for 7-day product consistency test (PCT) normalized B release, the effect of glass crystallization on glass durability can be predicted with an uncertainty less than 50% if the residual glass composition is within the range of the PCT model.

Effect of feed melting, temperature history, and minor component addition on spinel crystallization in high-level waste glass

Abstract Spinel crystallization affects the anticipated cost and risk of high-level waste (HLW) vitrification. Spinel, (Fe, Ni) (Fe,Cr)2O4, is the primary crystalline phase that precipitates from melts containing oxides of Fe, Ni, and Cr in sufficient concentrations. This study was undertaken to help design and verify mathematical models for a HLW glass melter in which spinel crystals precipitate and partially settle. To study melting reactions, we used a simulated HLW feed, prepared with co-precipitated Fe, Ni, Cr, and Mn hydroxides. Feed samples were heated up at a temperature-increase rate presumed to be close to that which the feed experiences in the HLW glass melter. The decomposition, melting, and dissolution of feed components (such as nitrates, carbonates, and quartz) and the formation of intermediate crystalline phases (spinel, sodalite, and Zr-containing minerals) were characterized using evolved gas analysis, volume-expansion measurement, optical microscopy, scanning electron microscopy (SEM), thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), and X-ray diffraction (XRD). Nitrates and quartz, the major feed components, converted to a glass-forming melt by 880°C. A nearly chromium-free primary spinel formed in the nitrate melt starting from 520°C and eventually dissolved in the borosilicate melt by 1060°C. Sodalite, a transient product of corundum dissolution, appeared above 600°C and eventually dissolved in glass. To investigate the effects of temperature history and minor components (Ru, Ag, and Cu) on the dissolution and growth of spinel crystals, samples were heated up to temperatures below or above liquidus temperature (TL) and then subjected to different (constant or cyclic) temperature histories and analyzed. The results show that the mass fraction of spinel as well as the composition and size of crystals depend on the chemical and physical makeup of the feed and the temperature history. Small crystals result from either preventing the primary spinel from dissolving in the glass-forming melt or from the presence of nucleation agents, such as RuO2 (additions of 0.06 mass% Ag2O and 0.06 mass% CuO did not help nucleate spinel). Cyclic temperature histories with the maximum temperature above TL and minimum temperature below TL kept spinel concentration far below the equilibrium concentration, especially when a nucleation agent was absent.

LIQUIDUS TEMPERATURE-COMPOSITION MODEL FOR MULTI-COMPONENT GLASSES IN THE Fe, Cr, Ni, AND Mn SPINEL PRIMARY PHASE FIELD

Abstract We developed an extensive T L database of simulated high-level waste (HLW) glasses within the spinel primary phase field. Partial-molar T L s, T i , were determined for all components that were systematically varied in database glasses ( i =Al, B, Ca, Cr, Fe, K, Li, Mg, Mn, Na, Ni, Si, Ti, U, and Zr). A clear relationship was found between the T i values and ion potential. This led to a new model that can accurately predict the T L of glasses within component concentration ranges of the database. The model gives slightly better predictability than a first-order expansion of T L in composition while using only half of the fitted parameters and offers an improvement in predictability over previously published models. The success of this model gives insight to the nature of component effects on T L , which warrants further investigation. Specifically, the concentrations of glass components appear to be influential on T L in proportion to the character of their bonds or their ionic potential.

Rhenium Solubility in Borosilicate Nuclear Waste Glass: Implications for the Processing and Immobilization of Technetium-99

The immobilization of technetium-99 ((99)Tc) in a suitable host matrix has proven to be a challenging task for researchers in the nuclear waste community around the world. In this context, the present work reports on the solubility and retention of rhenium, a nonradioactive surrogate for (99)Tc, in a sodium borosilicate glass. Glasses containing target Re concentrations from 0 to 10,000 ppm [by mass, added as KReO(4) (Re(7+))] were synthesized in vacuum-sealed quartz ampules to minimize the loss of Re from volatilization during melting at 1000 °C. The rhenium was found as Re(7+) in all of the glasses as observed by X-ray absorption near-edge structure. The solubility of Re in borosilicate glasses was determined to be ~3000 ppm (by mass) using inductively coupled plasma optical emission spectroscopy. At higher rhenium concentrations, additional rhenium was retained in the glasses as crystalline inclusions of alkali perrhenates detected with X-ray diffraction. Since (99)Tc concentrations in a glass waste form are predicted to be <10 ppm (by mass), these Re results implied that the solubility should not be a limiting factor in processing radioactive wastes, assuming Tc as Tc(7+) and similarities between Re(7+) and Tc(7+) behavior in this glass system.

The effect of experimental conditions and evaluation techniques on the alteration of low activity glasses by vapor hydration

Abstract The vapor hydration test (VHT) is used to study the reaction between waste forms and water with the aim to gain insight into their alteration behavior. In nuclear-waste immobilization, the test is primarily used as a screening tool to identify durable/non-durable glasses and as a convenient method to generate and identify alteration products for use in performance modeling. The lack of a standard procedure for conducting this test has resulted in a number of reported test methods, thus decreasing the ability to directly compare test results from different sources. To optimize the VHT procedure, a series of tests was conducted on simulated low activity waste (LAW) glasses at temperatures ranging from 150 to 300 °C with different volumes of water, specimen holders, specimen-preparation techniques, and data-evaluation methods. Reaction progress was monitored by measuring the thickness of the remaining glass and the alteration layer. The alteration rate at which glass is converted into the alteration products was determined by linear regression from the remaining glass thickness plotted as a function of time. The resulting procedure for conducting VHT eliminates problems associated with the measurement of alteration layers and enables direct comparison of alteration rates for different materials.

Glass Property Data and Models for Estimating High-Level Waste Glass Volume

This report describes recent efforts to develop glass property models that can be used to help estimate the volume of high-level waste (HLW) glass that will result from vitrification of Hanford tank waste. The compositions of acceptable and processable HLW glasses need to be optimized to minimize the waste-form volume and, hence, to save cost. A database of properties and associated compositions for simulated waste glasses was collected for developing property-composition models. This database, although not comprehensive, represents a large fraction of data on waste-glass compositions and properties that were available at the time of this report. Glass property-composition models were fit to subsets of the database for several key glass properties. These models apply to a significantly broader composition space than those previously publised. These models should be considered for interim use in calculating properties of Hanford waste glasses.

Liquidus temperatures of HLW glasses with zirconium-containing primary crystalline phases

Abstract Three empirical models are investigated for predicting liquidus temperatures ( T L ) as a function of glass composition for high-level waste (HLW) glasses with zirconium-containing primary crystalline phases. The three models are: (1) first-order multiple regression, (2) first-order single regression, and (3) second-order multiple regression. For each of two HLW data sets and their union, the models were developed using least squares regression on data from glasses with any zirconium-containing primary phase, and for the subset of glasses with zircon (ZrSiO 4 ) as the primary phase. The fitted models, evaluations of their performances, and interpretations of results are presented. The first- and second-order models fit very well for four of the six combinations of data set (A, B, A + B) and primary phase (zircon, zirconium-containing), and reasonably well for the other two combinations. For the composition regions and primary phases studied, T L is: (i) increased by ZrO 2 , Al 2 O 3 , and MgO, (ii) decreased by Na 2 O and Li 2 O, and (iii) moderately to negligibly affected by SiO 2 , B 2 O 3 , and CaO.

Dissolution and growth of spinel crystals in a borosilicate glass

The rate of dissolution and growth of settling crystals of spinel was measured optically in a borosilicate melt that was pre-heated at a temperature above liquidus to erase the effects of previous history. The Hixson–Crowell equation, which is based on Ficks first law, was used to determine mass-transfer coefficients (kH) for dissolution and growth; both were found to fit the same Arrhenius function of temperature (T). An attempt was made to estimate the diffusion coefficient (D) and the concentration-boundary-layer thickness (δ). The calculated values of δ compared well with experimental results and observations. The D vs. T function was similar to a literature function obtained for the dissolution of magnetite in sodium disilicate glass.

Database and Interim Glass Property Models for Hanford HLW and LAW Glasses

This report discusses a methodology for increasing the efficiency and decreasing the cost of vitrifying nuclear waste by optimizing waste-glass formulation. This methodology involves collecting and generating a property-composition database (for glass properties that determine waste-glass processability and acceptability) and relating these properties to glass composition via property-composition models. The report explains how the property-composition models are developed, fitted to data and evaluated, validated using additional data, used for glass-formulation optimization, and continuously updated in response to changes in waste-composition estimates and processing technologies. Further, the report describes a waste-glass property-composition database compiled from literature sources and presents the results from a critical evaluation and screening of the data for applicability to Hanford waste glasses. Finally, the report provides interim property-composition models for melt viscosity, liquidus temperature (with spinel and zircon primary crystalline phases), and Product Consistency Test normalized releases of B, Na, and Li. Models were fitted to a subset of the database deemed most relevant for the anticipated Hanford waste-glass composition region.